Today we’re joined by Elisa Lorello, a bestselling novelist and author of The AI Author Assistant.

ChatGPT made waves with its public release in November of 2022 and has ushered in a new era for research and writing. An incredible advancement in AI made available to the public, many praised it as a multi-purpose, catch-all solution to creativity in the workplace, while others feared it would do harm to artists and writers. Elisa will outline the benefits and limitations of AI, and explain how ChatGPT shouldn’t be feared, but instead embraced as a helpful tool for writers — in ways you may not expect.

Welcome to episode 76 of Published! Today, we're joined by Elisa Lorello, a best selling novelist and the author of the AI Author Assistant. Elisa will talk us through some of her revelations about ChatGPT and share how it can best serve writers.

Welcome to Published, a podcast by Greenleaf Book Group, where we'll discuss the ins and outs of the publishing industry, from writing a book and finding the right publisher to gearing up for a book launch. And now here's your host, Greenleaf Book Group CEO, Tanya Hall.

ChatGPT made waves with its public release in November of 2022 and has remained at the center of conversation since. An incredible advancement in AI made available to the public, many praised it as a multipurpose, catch-all solution to creativity in the workplace, while others feared it would do harm to artists and writers. No matter your perspective, anyone can agree that ChatGPT has ushered in a new era for research and writing.

Elisa will outline the benefits and limitations of AI and explain how ChatGPT shouldn't be feared, but instead embraced as a helpful tool for writers in ways you may not expect.

Tanya: Elisa, welcome to Published!

Elisa: Thank you for having me, great to be here.

Tanya: Yeah, I'm excited for this episode! So tell us a little bit about yourself and what you do.

Elisa: So I am an author. I've been writing full time since 2012 . Prior to that, I was a college writing instructor. So I've been doing novels since 2012. I also have a memoir and a writing source book, and then most recently I published this AI book (The AI Author Assistant).

Tanya: That's what we're here to talk about. Specifically today, AI, which is gonna be fascinating. It comes up, I would say daily for me in conversations with authors. So for those who are new to the topic, can you start by explaining in kind of layman's terms, what do we mean when we're talking about AI, and specifically ChatGPT since we'll be talking about that quite a bit today. And what are these tools designed to do?

Elisa: Well, I'm gonna talk about it strictly in the context of what it's doing for writers and what it's being used for for writers. In general, and my familiarity is mostly with ChatGPT so I'll also say that. It is what's known as a limited language model. Actually, it's not even limited anymore. It's an advanced language model. I actually asked ChatGPT to help me to define it because I'm not good at defining it.

Tanya: That's very, very meta.

Elisa: Very, very, very. Here's how it defined itself: It is designed to understand and generate human-like text based on the input it receives. Because it's called artificial intelligence, it has learned this language and learned data that has been given to it via prompts. Now, specifically, what it does in the writing world is: It can answer your questions, it can explain things to you, it can be used in similar ways as Google would, as a search engine to give you answers, and it can help you with problem solving. And for the writer themselves. It can help you with your writing process, and so that's what we could talk a little bit more about today.

Tanya: Yeah, that's what I'd love to dig into. So I think people are probably of two schools of thought, of course, that this is an amazing new tool or that this is a threat. So what would you say? Let's start with the best uses of ChatGPT for authors.

Elisa: I wanna acknowledge that, first of all. I am not 100 percent for, nor am I anti. So, what I'm really, really interested in is the ways to use this tool in responsible and ethical ways. I think what it really does and the way I wanted to use it is as an assistant.

So it can help you with things like brainstorming. So if you're looking for ideas for novels. Let's say you're interested in a particular genre, and you just want the basic information about it. So let's say I wanna start writing more hardcore contemporary romance. I think I my books already kind of fit that, but let's say I really, really wanna explore the genre and I say something like, so what are the most popular tropes of a contemporary romance? And I could ask the AI and ChatGPT would list the most popular tropes. So then I might take one of those tropes. So let's say office romance, you know, something like that. And I'll say let's brainstorm 10 ideas for an office romance. And then it will give me 10 ideas. And then from there, I might say, alright, give me some character names that are popular for contemporary romances.

You start using it as this brainstorming tool. What happened for me is that when I started playing with it in this way, my own creativity took over. So as it gives me these ideas for office romance, I'm reading through the ideas and I'm like, well, I don't really like any of those. But I like this little kernel from idea number 2 and this little kernel from idea number 8. So let me see how I could put those two together.

Or it gives me a name, you know, gives me character name ideas. And I'm like, well, I don't really, you know, well, I used Matthew in my last book, but maybe I'll use Matthews as the last name. In that way, it's good for stimulating my own creativity.

I also personally use it just for outlining because I was never ever good at that, and I never liked to outline at the onset of writing a novel. So now what I do is, if I have the premise already set up for novel idea, I'll give ChatGPT the premise and then I'll say "Give me a very general 3-act story outline and fill in beats here and there." And it gives me something very generic, but it gives me something to work with that I didn't have before. And again, my own creativity takes over.

Especially as the story develops for me in my own head, I could start plugging those into the outline, but the hard work is already done for me. It's always sapped my creativity to write an outline, and all of a sudden, this is now fueling my creativity. So those are the very, very basic ways in which I use it.

Now there are writers who are really diving in and using ChatGPT as a cowriter. And that is very heavily reliant on how you are prompting ChatGPT. There's a lot of practice and a lot of finesse to that.

I belong to a Facebook group of authors who use AI in various ways. And some of them really are, like I said, going all in and using it as a cowriter and generating 50,000 word novels from it. And it's not and the way they're using it to my understanding is it's not that ChatGPT is writing the whole novel for them and they're just kind of sitting back. It really is a co-writing experience.

So, like, for example, a few years ago, my husband and I co-wrote a novel. And so we we alternated the chapters — when I finished writing a chapter, I gave it to him, he knew where I was taking the story, he knew where to pick it up, and then, you know, he would bring it back to me and I would see what he wrote. That's how we did it. And then we would both of us would go in and tweak each other's dialogue and things like that. So that was the collaborative experience.

To my understanding, that's kind of the way they're using ChatGPT in a similar collaborative experience. They may prompt it to write a particular scene and give it all the details of the setting and the characters and what they want to see happen and ChatGPT will write that scene, and then they go and tweak it, and then they pick up where that scene left off, and then they feed ChatGPT the next scene, and so it really is a back and forth.

Now this, I know, is where the controversy comes in. The publishing world and other authors are saying that's plagiarism, there are intellectual property, concerns, especially if you're inputting your own novel to do that. You're feeding it. You know, you're kind of feeding the beast and now they have all of these words. And they have your story in there. And so somebody else might prompt something and now your story's coming out of that. That's the concern.

Tanya: Yeah. That's so interesting. You know, I think even if somebody is turned off at the notion of ChatGPT, it's really good to get in there and just play with it. And I did this myself because I thought, okay. How could it be helpful?

Anyone who's listened to this podcast knows I have a book called Ideas, Influence, and Income. Titling a book is a headache. And it's so important, but I think, like you said, sometimes you can just kinda hit a wall and it saps your creativity.

So, the books's already out, of course. It already has a title, but I thought, "I wonder how it would have done if I told it what the book was about, and then prompted it to give me a list of titles." And it was somewhat validating because its titles were awful. And then I told it what the actual title was and thought, "Could you give me some alternate titles?" And it gave me something like, "Content, Community, and Commerce." And I was like, well, it kept the alliteration, but that's still awful.

Elisa: Yes. Yeah. That was my experience with it too. Every time I've tried to brainstorm titles, they're really bad. The paid version of ChatGPT, which I believe is ChatGPT 4, from what I've heard is a little bit more sophisticated, and I think you'll get a better sampling.

TanyaL Yeah. That's interesting. But, you know, one of the other things that I learned by playing with it is that you could tweak the outcome of the writing by instructing it to write in the voice of another writer, which is also very controversial. Right? So I went over the top and told it to write something in the voice of John Updike. And again, I thought it was garbage what it turned out, but it was kind of in the same vein. It's just super, you know, overly descriptive. In my opinion, but ChatGPT isn't John Updike, didn't quite nail it.

But those prompts are so important, I think is what those of us who have tinkered with this tool have come to learn. Can you speak about how do you get to the right prompt?

Elisa: That's a great question and that really is the key to making it work for you. And I think it becomes a trial and error. Be as specific as possible when you're directing it. Depending on what you're using it for, for example, if you are using it as a cowriter, then you have to direct it in terms of cadence and in terms of syntax, and voice and things like that.

Now talking about style, I did the same thing when I was playing around with it. Asking it to give me the style of another author or favorite author or something like that. What I've since learned is if you are embracing it as co-author and using it in more serious ways and not really just playing around and tinkering, it's advised that you don't do that. Because of, again, these intellectual property issues.

If I were to do this, and I'm not saying I've done it seriously, if I was gonna co write a novel and I said, "I want you to write this in the style of Janet Evanovich." I said, you know, something like that. Then I use that, and copy and paste it and then I wind up publishing that novel. I could be in some trouble with that. Don't specifically reference any author.

I mean, if you say "in the style of a contemporary romance writer" — I know that's not what Janet Evanovich writes, I'm going back to my original example of genre. You're still probably in a little bit of of tricky water, but not as much as if you name a particular author.

Of course, you're completely inbounds, I would be completely inbounds to say, "Write it in the style of Elisa Lorello," which again, I've tried, and that was hit or miss. Even just saying write it in my style was a little bit, which is something else by the way that that I'm gonna segue into:

Don't overly rely on ChatGPT for accuracy. That goes for research, too. It is not always accurate. I mean, for example, like my husband was playing around with it, and he said, "Write me a a biography for author Craig Lancaster," and it got most of it wrong. I did it myself.I said, rank Elisa Lorello's books from the best selling to the least best selling. And it only listed 5 and 2 of the 5 weren't mine, and the top 1 was my worst selling book.

In more serious things, I was trying to research a legal issue, and it gave me very inaccurate information. I kinda got into an argument with it. And I'm like, no, this isn't this isn't the outcome and it doubled down and gave me links to websites that didn't exist, and quoted press releases that didn't exist. So I had to go and try to fact check all that anyway.

ChatGPT does save you a lot of time, but you have to be careful because like I said, it's not always accurate. But I do use it occasionally in lieu of Google when I want to know something. Like, for example, I wanted to know, it was really almost like just a scene where this character is thinking about opening a business in a particular location. And I just kinda asked it, what would a person do if they wanna open up this kind of business in this kind of location? I wasn't worried about a hundred percent accuracy. I just wanted the bare basics of it.

Tanya: So trust but verify.

Elisa: Exactly. Yeah. Which you have to do with everything anyway.

Tanya: Hundred percent. Yes. You should be doing that. In the news, lately, I'm sure anybody who's remotely involved in this industry sees the coverage of — Is ChatGPT going into replace writers, ghostwriters, copy editors, the list goes on and on. And you hear stories of companies that are even cutting the ranks in terms of their creatives, saying that we've replaced it now with AI. So do you think that writers should perceive AI as a threat?

Elisa: There is disruption going on. There is absolutely disruption going on in the industry and everything you said, yes, is happening. And I'm especially concerned for the screen writers because that seems to be a really big issue and one of the things they're on strike about is this kind of overall threat. I don't want to be cavalier about that.

On the other hand, I think in a lot of ways there's been some overreaction, and I'm speaking in a general term. So I'm not speaking in these very, very specific things that you had mentioned. Yes, those things are absolutely happening. And, yes, that's not good for writers or creatives as a whole. So I don't wanna minimize that.

In the big picture and in the general sense, I don't think it's going to do the damage that some people are saying it's going to do, and especially not long term. And one of the reasons why... and I could just be an optimist, or I could just be naive, I'm not sure what the answer is to that.

When when you see what it generates, there is a human element missing from its prose. There is almost like a dullness to it, and it's just lacking a soul. I just don't see how you can produce any kind of quality content. And I'm talking in terms of the arts. I'm not necessarily talking about technical manuals and things like that. And I've actually really struggled using it as a cowriter, mostly because I just feel like it's missing that human element. I feel like it's missing an emotional component. I almost wanna say it's spiritual.

And especially, I just know myself as a novelist. So much of what I'm bringing to the novel is borne of my own human experience. It comes from memory. It comes from something I witnessed. It comes from something I felt very deeply about at the time. It comes from something that I've explored over and over again thematically. I don't think that's any kind of language model is going to be able to replicate that.

Readers are gonna pick that up because that's the whole reason why a reader is picking up a novel in the first place: They want the connection.

Tanya: We're kind of hitting on all of the touch points for concerns with ChatGPT, but I think an important one that you touched on earlier is plagiarism. So there's two layers to it. You gave an example of how you're feeding it your work, and who knows how it may use that in the future. But there's another layer, which is when an author is working with a publisher, they're representing that the work that they are submitting is their original work and has not been plagiarized.

So is there a chance that ChatGPT is plagiarizing on your behalf if you're using it in this cowriting capacity? Can you speak to how an author can protect themselves against these potential areas of trouble?

Elisa: You just gotta do your own homework and find out what's the latest on intellectual property copyright with AI because they're going to have to really start regulating it, or if not regulating it, clearly define it.

I think publishers right now, because it's a little murky right now, most publishers aren't going near it. They don't want AI-assisted writing. I mean, if I was ever gonna do it, if I was ever gonna write a co-author AI novel, I would self publish it. I don't even think I would submit it to a publisher. And I've even wrestled with, would I tell a reader that it's been co-authored with AI? And the fact that I have to even ask that question tells me maybe I shouldn't.

Tanya: Exactly.

Elisa: So that's why I've not gone that far. Like I said, mostly what I've been using it for is writing outlines and brainstorming and writing, it makes my schedule. Sometimes it's just like literally a personal assistant where it's like, I need I have this to-do list today and put it in order for me and give me some scheduled times to do it, and it's great for that. It's been really, really good for my productivity in that sense.

Tanya: That's interesting. So supposing some of our listeners are authors who have not yet tinkered with AI and and ChatGPT probably specifically, where should they start?

Elisa: Start with playing. Start with that kind of brainstorming. The pleasant surprise was it wasn't taking my creativity, it was fueling it because everything I saw, I either said I could do it better and I did, or it gave me it was a springboard for me where I said, I don't like that, but oh my god, I just got this idea. And now I'm really excited about this idea.

I really do think there are responsible and ethical use for it. And I said that even in the little book that I wrote that that's my goal from the get go. I just want to find positive, responsible ways to use this. Because it's not going away. And yes, there will be disruption. So while this disruption is going on, let us steer the ship rather than it steering us.

Tanya: I love it. And you've given us so much great insight and advice today. Would you like to tell us the name of this little AI book?

Elisa: Yes. That would be helpful, wouldn't it? So it's called The AI Author Assistant, and I was very specific in that title. And, yes, ChatGPT did help me brainstorm that title.And I was very specific about calling it an assistant. So I think the original title it gave me was The AI Author, and I didn't want that again to have the connotation that it does everything for you and it's writing all your books for you and I don't advocate for that. But I do advocate for using it as as an assistant. That is always always always my number 1 goal.

And yes, a portion of that book is written by ChatGPT, but I'm very, very upfront with that, and I kind of almost did it as a dialogue. So what ChatGPT has to say is in its own font, and then what I have to say is in my own font. And you could very, very clearly see the stylistic difference between the 2. There were times when I literally quoted it, so it was like a source. So I was very, very up front about the ways I used it. And I used it a lot to demonstrate it. But, yes, it's called the AI Author Assistant. You'll read it in an hour.

Tanya: It's a field guide.

Elisa: A field guide. It's a field guide.

Tanya: Yes. Awesome. Well, you've given us so much good stuff to think about today, Elisa. Thank you so much for joining us.

Elisa: Thank you so much. I had a great time.

That's it for our episode today with Elisa. We hope you enjoyed learning about ChatGPT and how to use AI as a tool for your writing.

For notes and resources from today's show, go to greenleafbookgroup.com/episode76.

You can also find advice for writing, publishing, and promoting your work in my book, Ideas, Influence, and Income, which you can learn more about at ideasinfluenceandincome.com.

If you've enjoyed our show, please rate and review us on iTunes, Spotify, or wherever it is you listen to your podcasts.It means a lot to have your feedback and helps us make sure we're answering your publishing questions.

A big thank you Eleanor Fishbourne and Madison Johnson, who produce the Published podcast, and we will be back next month with another episode.

Elisa Lorello is the bestselling author of twelve novels and one memoir. The youngest of seven, she grew up on Long Island and graduated with two degrees from University of Massachusetts-Dartmouth. Since 2010 she’s sold over a half-million units worldwide and has been featured in the Charlotte Observer, Woman’s World magazine, Rachel Ray Every Day magazine, The Montana Quarterly magazine, Writer’s Digest Online, and Jane Friedman’s Five On blog series.

Elisa's book The AI Author Assistant is available on Amazon.